[GE users] only 1 set of Qs will run

reuti reuti at staff.uni-marburg.de
Thu Nov 27 18:46:54 GMT 2008

Am 27.11.2008 um 17:08 schrieb Harry Mangalam:

> Thanks for the suggestions, John,
> On Wednesday 26 November 2008, John Hearns wrote:
>> Harry,
>> on each type of node run 'mount' and look at exactly how the
>> SGE_ROOT filesystem is mounted.
> one every node, the (anonymized) mount line is:
> mount_host:/share/sge62 on /sge62 type nfs
> (rw,nosuid,intr,hard,timeo=20,rsize=8192,wsize=8192,addr=x.x.x.x)
>> On the AMD64 nodes, consider using local spooling - you can find
>> out how to do that in the NFS_Reduce HOWTO. Basically create a
>> directory called (let's say)
>> /var/spool/sge on the remote host, make it owned by the sge user.
>> Define sge_execd_spool = /var/spool/sge  for that machine. Restart
>> sgeexecd.
> I had already done this (probably lost in the long initial post) -
> that's why I restarted the sge_execd on the nodes and qmaster daemon.
> And afterwards the NFS errors stopped, tho the inability to run jobs
> on those Qs remains.  All AMD nodes are now spooling to a local dir
> (/tmp/sge_execspool/).

Small add on: I would also suggest to put it in /var/spool/sge as  
John did. If you want to clean /tmp from time to time, you always  
have to take care not to delete this directory by accident.

-- Reuti

> -- 
> Harry Mangalam - Research Computing, NACS, E2148, Engineering Gateway,
> UC Irvine 92697  949 824-0084(o), 949 285-4487(c)
> ---
> Good judgment comes from experience;
> Experience comes from bad judgment. [F. Brooks.]
> ------------------------------------------------------
> http://gridengine.sunsource.net/ds/viewMessage.do? 
> dsForumId=38&dsMessageId=90107
> To unsubscribe from this discussion, e-mail: [users- 
> unsubscribe at gridengine.sunsource.net].


To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].

More information about the gridengine-users mailing list